Hormonal regulation of non-CFTR ion channels in the endocervix Ion channels are hormonally regulated

Abstract

Objective:

To characterize ion channel expression and localization in the endocervix under different hormonal conditions using a non-human primate (NHP) primary endocervical epithelial cell model.

Design:

Experimental.

Setting:

University-based, translational science laboratory.

Interventions:

We cultured and treated conditionally reprogrammed primary rhesus macaque endocervix cells with estradiol and progesterone and measured gene expression changes for several known ion channel and ion channel regulators of mucus secreting epithelia. Using both rhesus macaque endocervical samples and human samples, we localized channels in the endocervix using immunohistochemistry (IHC).

Main Outcome Measures:

Relative abundance of transcripts was evaluated using real-time PCR. Immunostaining results evaluated qualitatively.

Results:

Compared to controls, we found that estradiol increased gene expression for ANO6, NKCC1, CLCA1 and PDE4D. Progesterone downregulated gene expression for ANO6, SCNN1A, SCNN1B, NKCC1, and PDE4D (P ≤ .05). IHC confirmed endocervical cell membrane localization of ANO1, ANO6, KCNN4, LRR8CA, and NKCC1.

Conclusion:

We find several ion channels and ion channel regulators that are hormonally sensitive in the endocervix. These channels therefore may play a role in the cyclic fertility changes in the endocervix and could be further investigated as targets for future fertility and contraceptive studies.

Capsule:

Sex-steroids demonstrate selective transcriptional regulation of ion channels in the endocervix of the non-human primate.

Introduction

The cervix produces specialized mucus secretions from the glandular epithelial cells of the endocervix (1,2). This mucus is composed of water (90 – 98%), proteins, lipids, carbohydrates, nucleic acids, and electrolytes (3). The most prominent organic material are large mucin glycoproteins which serve as the framework for the mucus gel as these large oligomers form networks of water-binding glycan chains (3). While mucins provide the structure for hydration, the amount of hydration is regulated by the electrostatic and pH environment in the lumen (4). Thus, epithelial ion channels are major determinants of mucus hydration and viscosity as they determine both the transcellular movement of solutes and water, as well as mucin conformation and water binding capacity (3).

Within this physiology of mucus hydration and epithelial ion transport, we do not completely understand how it may play a role in cyclic mucus changes. We know that mucus grossly changes dramatically throughout the menstrual cycle. For most time points during the menstrual cycle, mucus is viscous, scant, and cellular. However, during the peri-ovulatory window, mucus becomes thin, watery, a-cellular, and highly permissive to sperm (1, 5). We know that these changes are estradiol and progesterone dependent (6). However, it is not well known which critical biological mediators are responsible for this shift. Elucidation of the physiology and regulatory pathways of mucus secretion in the cervix would have implications for the way we diagnose and treat infertility, as well as our ability to drug the cervix non-hormonally for contraceptive development.

To date, the role of ion channels in mucosal epithelia has been most extensively defined in the airway (7). The most well-described ion channel is the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR), an epithelial anion channel in which a mutation causes the disease Cystic Fibrosis (CF). In CF, decreased chloride (Cl⁻) and bicarbonate (HCO₃⁻) secretions from defective CFTR result in abnormal, thickened secretions leading to multi-organ dysfunction. However, the landscape of ion channels and their regulators is complex extending far beyond CFTR. Additional anion channels such as calcium-activated chloride channels (CaCCs), chloride channel protein 2 (CLCN2), volume-regulated anion channels (VRACs), and cation channels such as the epithelial sodium channel (ENaC), sodium potassium chloride cotransporter (NKCC), and calcium-activated potassium channel (KCNN) work together in healthy mucosal systems to produce electrostatic environments optimal for biological function. These channels are additionally regulated and their function is further augmented with the assistance of accessory proteins and enzymes such as calcium-dependent chloride conductors (CLCAs), and phosphodiesterase inhibitor enzymes such as PDE4 (8, 9).

In these studies, we sought to investigate whether several important non-CFTR ion channels found in non-reproductive mucosal systems are present in the endocervix and whether their expression is hormonally regulated by E2 and P4. In order to measure gene expression changes specific to the endocervical epithelium, we performed experiments using an established primary endocervical cell culture derived from Rhesus macaque (Macaca mulatta). We previously demonstrated that these cultures expand robustly, maintain expression of hormone receptors, and secrete mucus that is similar to what occurs in vivo (10-12). We use the nonhuman primate because of its homologous endocervical anatomy including a glandular mucus-producing epithelium and similar female hormone physiology, as only primates have menstrual cycles.

Methods:

Primary Cell Generation and Culture

The Oregon National Primate Research Center (ONPRC) provided care and husbandry for the macaques in this study. All animal study procedures were reviewed and approved by the ONPRC Institutional Animal Care and Use Committee (IACUC).

Conditionally Reprogrammed Endocervical Cells

We previously described using conditional reprogramming to produce primary endocervical cell cultures (CRECs) from reproductive-aged female Rhesus macaques (Macaca mulatta) (10). Briefly, we obtained cervical tissue from animals undergoing necropsy for reasons unrelated to this study. We scalpel isolated the endocervical epithelial tissues and digested them before plating them on collagen-coated plasticware. All cells in this study were incubated at 37°C and 5% CO₂. We cultured cells in expansion media [373 mL complete DMEM (DMEM [(500 mL) of DMEM with L-glutamine (Fisher, Waltham, MA, USA), FBS (10%, ATCC), penicillin/streptomycin (1 x 104 units/mL, Fisher)].), insulin (5 μg/mL, Sigma Aldrich, St. Louis, MO, USA), amphotericin B (250 ng/mL, Fisher), gentamicin (10 μg/mL, Gibco, Waltham, MA, USA), cholera toxin (0.1 nM, Sigma Aldrich), epidermal growth factor (0.125 ng/mL, Life Technologies, Carlsbad, CA, USA), hydrocortisone (25 ng/mL, Sigma Aldrich), and ROCK inhibitor Y-27632 (10 μM, Enzo, Farmingdale, NY, USA)] and followed previously reported methodologies of conditionally reprogrammed cell cultures to propagate and expand the primary endocervical cells (2, 10). During initial plating after tissue digestion, any samples found to have native tissue fibroblast outgrowth were not utilized for the course of this study. Briefly, cells were co-cultured with irradiated 3T3-J2 swiss mouse fibroblast cells (Kerafast, Boston, MA, USA). During expansion and propagation, samples were double-trypsinized to remove fibroblasts prior to replating the epithelial cells. To differentiate cultured cells, we seeded 1 x 10⁵ cells on permeable supports (Costar Transwell, Corning, Action, MA, USA, 12mm diameter, 0.4μM pore size) in phenol-free, calcium (Ca⁺²) supplemented (total Ca⁺² concentration = 0.4 mM, Sigma Aldrich) commercially available media (ReproLife CX, Lifeline Cell Technology, San Diego, CA, USA).

Hormone Treatments

We treated differentiated, primary endocervical cells with different hormonal conditions to mimic mid-cycle and mid-luteal conditions (10, 13, 14). For mid-cycle conditions, we added 17β-estradiol (E2) alone (10⁻8 M, Sigma Aldrich) to differentiation media once daily for nine days. For mid-luteal conditions, we initially primed cells with E2 (10⁻⁸ M) for seven days, then followed this with progesterone (P4) (10⁻⁷ M, Sigma Aldrich) and E2 (10⁻⁹ M) for 48 hrs. Additionally, we included an E2 withdrawal condition where cells were cultured for 7 days with E2 (10⁻⁸ M) followed by 48 hours without any hormones. The inclusion of this non-physiologic condition was conducted in order to compare P4 independently to the withdrawal of E2 in order to gain insight into which hormones may be driving regulatory changes. Finally, we also included a no-hormone control culture for nine days.

RNA Isolation and cDNA Synthesis

Following the hormone treatment, we extracted total RNA from the cell cultures using the TRIzol RNA+ (Invitrogen, Gaithersburg, MD, USA) mini-kit following the manufacturer’s instructions. RNA concentration and purity were measured using a Nanodrop system (Thermofisher, Waltham, MA, USA). We used the Superscript III First Strand cDNA Synthesis kit (Invitrogen) following the manufacturer protocol to generate cDNA from 1μg of RNA. The protocol includes DNase I treatment prior to the reverse-transcriptase reaction.

Gene Expression Analysis

We used the Fluidigm BioMark HD Real-Time PCR system to perform relative quantification of transcripts for target genes according to manufacturer instructions (Fluidigm, San Francisco, CA, USA). Primer sequences were designed to span exon-exon junctions (Supplementary Table 1) and each gene target was previously validated for real-time PCR utilizing control tissues. In addition to gene targets for ion channels, we also evaluated gene expression of progesterone receptor (PGR) and estrogen receptor (ESR1) as controls given their known gene expression changes under the influence of E2 and P4 (16). We loaded assays and samples, both in duplicate, into a 96.96 Dynamic Array Integrated Fluidic Circuit (IFC) Chip (Fluidigm). We analyzed a total of n=3 primary cell lines with n=4 experiments per cell line.

Data was collected by BioMark Data Collection software (Fluidigm) and analyzed using qBase+ (Biogazelle, Zwijnaarde, Oost-Vlaanderen, Belgium). Target genes were normalized using reference genes determined by geNORM to be most stable (17). Our final housekeeping genes consisted of a panel of HMBS and RPL32. We assessed mean differences in relative abundance of transcripts using an ANOVA Turkey Post-hoc test with relative abundance of transcripts as the outcome and treatment as the predictor with E2 as the referent. All analyses were conducted in Stata version 15.1 (Stata Corp, College Station, TX, USA).

Immunohistochemistry

In order to conserve animal resources, we used archived samples of paraffin-embedded rhesus macaque mid-cervical mucosa for immunohistochemistry (IHC). These samples were obtained from adult oophorectomized animals that were hormonally treated with hormone implants as previously described (18). Tissues used in this study were from animals that either had an E2 implant only or had both an E2 and P4 implant. Lung and kidney tissues were used as positive control tissue. Additionally, for evaluation of ion channels with no validated rhesus antibody, we used histologic sections of human endocervix from deidentified archived tissue obtained from pre-menopausal women undergoing loop electrosurgical excision procedure (LEEP) for ectocervical dysplasia with benign endocervical findings. These samples do not have data regarding menstrual cycle timing. IHC using ion channel antibodies was diluted according to company recommendations (Supplemental Table 2) and conducted on 5μm paraffin sections using a standardized technique (19). Digital images were taken using a VisionTek Digital Microscope system (Sakura Finetek, Torrence, CA, USA).

Results

Gene Expression changes of Estrogen and Progesterone Receptor

We measured changes in PGR and ESR1 in order to assess hormonal sensitivity in our cell cultures (Figure 1). We observed PGR expression to be induced by E2 (EE) and down-regulated by the addition of P4 (EP). We found ESR1 gene expression to be present at baseline, but also significantly down-regulated by P4. Additionally, E2 withdrawal led to decreases in both ESR1 and PGR gene expression.

Figure 1.

Relative transcript abundance of steroidal hormone receptors in NHP CRECs A) PGR B) ESR1 (ANOVA, p < 0.05, ** ≤.01, *** ≤ .001)

Anion channels in the endocervix

We measured gene expression changes for two members of the CaCC family: ANO1 and ANO6, as well as CLCN2 and a volume-regulated anion channel, LRR8CA. For ANO1 gene expression, we did not see significant differences between any of our conditions (Figure 2A). Using sections of rhesus and human endocervix, we observed immunostaining of ANO1 at the apical membrane (3A-C). We found ANO6 gene expression to be up-regulated under E2 treatment, and down-regulated with P4 (Figure 2B). ANO6 localization appeared apically for tissues treated (Fig. 3D-F). We did not see significant differences in CLCN2 gene expression, and commercially available CLCN2 antibodies were not sufficient for staining rhesus tissues. LRR8CA gene expression did not demonstrate significant differences between conditions (Figure 2D), and protein expression appeared along the membrane and within the cytosol as described in other studies (20, 21). (Figure 3G-I).

Figure 2.

Relative transcript abundance of ion channels in NHP CRECs including A) ANO1 B) ANO6 C) CLCN2 and D) LRRC8A. (ANOVA, p < 0.05, ** ≤.01, *** ≤ .001)

Figure 3.

Histologic sections of rhesus macaque endocervix and deidentified human endocervix stained for ANO1, ANO6, and LRR8CA. All bars = 80μm. E2 represents E2 only implant. P4 represents E2 and P4 implant. ANO1 and ANO6 localization appeared along the apical membrane. LRR8CA localization appeared along the apical membrane and cytosolically.

Cation channels in the endocervix.

We evaluated three different epithelial cation channels genes, including two subunits of ENaC (SCNN1A, and SCNN1B), potassium calcium-activated channel subfamily N member 4 (KCNN4), and sodium potassium chloride transporter 1 (NKCC1). In both ENaC subunit genes (SCNN1A, SCNN1B), P4 down-regulated gene expression (Figure 4A-B). Commercially available ENaC antibodies were not sufficient for staining rhesus tissues. We did not observe hormonal transcriptional regulation of KCNN4 (Figure 4C). KCNN4 localization appeared along the basal membrane and cytosolically as expected (Figure 5A-C). NKCC1 gene expression was up-regulated by E2 conditions and down-regulated in P4 conditions (Figure 4D). NKCC1 localization appeared along the basal membrane along with nuclear staining as expected (Figure 5D-F).

Figure 4.

Relative transcript abundance of cation channels in NHP CRECs including A) SCNN1A B) SCNN1B C) KCNN4 and D) NKCC1. (ANOVA, p < 0.05, ** ≤.01, *** ≤ .001)

Figure 5.

Histologic sections of hormone-controlled rhesus macaque endocervix and deidentified human endocervix stained for KCNN4 and NKCC1. All bars = 80μm. E2 represents E2 only implant. P4 represents E2 and P4 implant. KCNN4 localization appeared cytosolically and along the basal membrane (A-C). NKCC1 localization appeared along the basal membrane, and within the nucleus (D-F).

Additional Ion Channel associated genes

Finally, we evaluated two different accessory proteins, CLCA1 and PDE4D. CLCA1 assists and modifies anoctamin ion channel function by targeting and activating CaCCs. PDE4D assists in cAMP degradation, affecting cAMP-dependent CFTR’s ability to actively utilize cAMP in Cl⁻ secretion. Our results demonstrated that E2 up-regulates CLCA1(Figure 6A). Similarly, for PDE4D, we found that E2 up-regulates gene expression compared to the no-hormone control, and P4 subsequently down-regulated PDE4D gene expression compared to EE conditions (Figure 6B).

Figure 6.

A) Relative transcript abundance of CLCA1 in NHP CRECs (ANOVA, p < 0.05) B) Relative transcript abundance of PDE4D in NHP CRECs (ANOVA, p <0.05) * ≥ .05, ** ≥ .01, *** ≥ .001

Discussion

The ion channels and associated proteins we investigated in this study are highly conserved and play key roles in regulating the characteristics of secreted mucus in other mucosal tissues (7, 22). However, the role these channels play in the cervix has yet to be defined. Secreted mucus varies throughout the menstrual cycle in response to E2 and P4. Thus, measuring the hormonal sensitivity of these channels may provide insight into which channels may play a role in hormonal mucus changes. Our study leverages a novel in vitro primary cell model from NHPs, allowing us to perform experiments corresponding to differing hormonal states in the lower reproductive tract, as well as non-physiological conditions in menstruating individuals. We found that several ion channels demonstrate transcriptional regulation by E2 and P4, suggesting that these channels could play a role in observed cervical mucus changes during the menstrual cycle.

We previously demonstrated that CFTR is a cyclically expressed anion channel in the endocervix (11). However, CFTR is only one channel in a larger ecosystem of ion channels that regulate extracellular secretions in mucus-producing cells (Figure 7). Non-CFTR channels, anion channels in particular, may offer compensatory and complementary action. Members of the anoctamin family (ANO1, ANO6) are calcium-activated chloride channels (CaCC) and have been of great interest in CF research as they also shuttle Cl⁻, potentially serving as an alternate route for Cl⁻ and its associated water transport (23). Similarly, CLCN2 (a voltage-gated Cl⁻ channel), and the volume-regulated ion channel (LRR8CA) both contribute to Cl⁻ secretions in mucosal systems (10, 24-27). In our studies only ANO6 saw significant hormonal regulation.

Figure 7.

Estradiol (E2) and progesterone (P4) bind their receptors in the cytosol before translocating to the nucleus where they transcriptionally regulate mucus characteristics including synthesis of mucins, enzymes, ion channels, and other mucus-associated genes. Apical anion channels ANO1, ANO6, CFTR, and LRR8CA primarily move Cl⁻ into the lumen, driving water secretion through primarily paracellular pathways. CLCA1 is a putative regulator of the anoctamin family channels including ANO1, and ANO6. Members of the PDE4 family such as PDE4D decrease cAMP availability, in turn modulating the cAMP-dependent CFTR channel. The anion channels, particularly CFTR are counterbalanced by apical cation channels such as ENaC, which move Na⁺ intracellularly. Together these channels regulate the hydration of mucins in the endocervical lumen. Meanwhile, basolateral ion channels, CLCN2, KCCN4, and NKCC1 maintain electrostatic gradients within the epithelia. Reduced NKCC1 function diminishes CFTR Cl⁻ transport. Figure created with BioRender.com

Cation channels often act as counter-regulatory elements to anion channels. ENaC is comprised of three subunits, produced by the SCNN1A, SCNN1B, and SCNN1G genes, together they move sodium (Na⁺) intracellularly. Thus, when CFTR is active, ENaC is usually inactive, moving electrolytes and water net outward. Conversely, an active ENaC along with inactive CFTR drives mucus dehydration and may exacerbate channel mutation phenotypes (28). It has been hypothesized that inhibition of ENaC can compensate for some of the mucus changes seen in CF by reducing Na⁺ and water resorption with dysfunctional Cl⁻ secretion (29). ENaC expression has been previously described in the mouse and rodent cervix (30, 31). While some of these previous studies demonstrate up-regulation during P4 dominant phases (31, 32), we observed ENaC down-regulation under the addition of P4 in our cell culture. Of note, other studies were not performed in the NHP model. Wherein NHPs are cyclically similar to that of humans and menstruate.

Basolateral cation channels drive electrostatic gradients intracellularly and impact apical ion function. For instance, KCNN4 must be functional in order for apical Na⁺ absorption to occur and may play a role in Cl⁻ secretion, as channel inhibition reduces conductance (33) and activation stimulates secretion (34). NKCC1 is a basolateral channel and in part responsible for the uptake of Cl⁻ across the basal membrane, and maintains potential differences across the cell interior through potassium (K⁺) (35). Blunting NKCC1 appears to alter CFTR-dependent Cl⁻ transport and secretory output (36, 37). Our results found NKCC1 expression to be up-regulated by E2, and down-regulated by P4. We additionally describe two accessory proteins that support ion channel conductance that are regulated hormonally: CLCA1 and PDE4D. CLCA1 is a channel accessory protein and appears to target and activate CaCCs (8). PDE4 inhibitors have been shown to restore CFTR activity in human airway epithelial cells in vitro (9) through increased cAMP availability for the cAMP-dependent CFTR channel. Taken together, our studies indicate that sex steroids globally increase transcription of a swath of ion channels suggesting that the fluctuating mucus changes are the result of large-scale regulation of ion channel behavior and not simply that of a single channel.

Our studies use NHPs, a uniquely relevant model for studying the endocervix as their hormone cyclicity and endocervical anatomy are similar to human physiology. Our group has previously established that reprogrammed primary cell culture lines derived from macaque endocervical tissues maintain their phenotype, are hormone sensitive and are able to generate physiologically relevant mucus similar in protein composition to human mucus (11, 12). Moreover, utilization of a primary cell culture allows us to investigate regulation specific to epithelial cells. Previous gene expression studies use whole organ cervix or endocervical brushings, which may be confounded by non-epithelial cell populations (38).

We utilized hormonal conditions that replicate mid-cycle E2 and luteal phase P4. We recognize that in vitro conditions may not fully demonstrate the changes seen in vivo. Our conditions lack stromal-epithelial interactions and other physiologic conditions such as the lack of an LH surge or the presence of a corpus luteum. Several of our selected ion channels did not demonstrate transcriptional change under in vitro hormonal treatment. One possibility is that while these channels may influence mucus secretions, they do not play a role in cyclic mucus fluctuations in the cervix. Another possibility is that their regulation does not occur through gene transcription, but instead another mechanism such as channel availability or channel function. In this study, we did not measure the impact of gene expression on mucus secretion itself. In vitro measurements of mucus hydration are difficult, but can be done using surrogates such as air surface liquid and rheology (39, 40). Mucus regulation itself is complex and there are likely many other contributors to mucus changes in the menstrual cycle. Other candidate arbitrators of mucus characteristics that have shown hormonal fluctuation include enzymes for mucin glycosylation (41), as well as proteases, sialidases and sulphatases (42).

Conclusion

In summary, our studies survey hormonally regulated ion channel expression of non-CFTR ion channels in the endocervix. This information provides new data regarding cellular mechanisms in the endocervix that may influence mucus hydration, and therefore provide possible targets for future fertility and contraceptive studies.

Attestation statement:

• The subjects in this trial have not concomitantly been involved in other randomized trials (If applicable).

• Data regarding any of the subjects in the study has not been previously published unless specified

• Data will be made available to the editors of the journal for review or query upon request